Venture Debt Explained: How Startups Borrow

Introduction

Ferrovane, an invented Series A software company, has $12 million in the bank and burns $1 million a month: twelve months of runway. Its Series B is about a year away, but the milestone that would earn a strong Series B price, roughly $10 million of annual recurring revenue, looks closer to eighteen months out.

The board has two ways to close the gap. Existing investors could put in another $10 million of equity at a flat price, leaving every shareholder with about a tenth less of the company. Or Ferrovane could borrow $10 million from a specialist lender, repay it over three years, and hand over a small warrant instead of a block of new shares.

The second route is venture debt: secured loans to companies with no profits and no realistic way to repay except by raising more equity. It looks reckless until you see what the lender is underwriting, which is where the explanation starts, before loan terms, warrants, covenants, the lenders after Silicon Valley Bank's 2023 failure, and a worked comparison of dilution against debt cost.

Lender and investor can fund the same company in the same month and want different things: the equity investor needs a large outcome, while the lender mostly needs the company to survive long enough to raise again.

Equity is expensive in good outcomes and cheap in bad ones; debt is exactly the reverse.

Who Borrows and When

The typical borrower is a venture-backed company between equity rounds, ideally one that has just closed a round, because that fresh equity is what pays the interest. A National Bureau of Economic Research (NBER) working paper by Jesse Davis, Adair Morse and Xinxin Wang found venture debt in roughly a third of early-stage financing rounds in their data. Typical uses are extending runway to a milestone, holding an undrawn tranche against a late round, funding equipment, or bridging to a planned sale or IPO.

Venture Debt

Venture debt is a senior secured loan to a venture-backed company that is not yet profitable, usually made alongside or soon after an equity round. It typically runs two to five years, starts with an interest-only period, carries a floating rate plus fees, and comes with a small warrant. Lenders underwrite it mainly on the company's equity investors and the likelihood of a further round, not on current cash flow.

Lending Against the Next Round

A conventional lender asks whether cash flow can service the debt. A venture lender knows it cannot, so it asks whether the owners will keep funding the company. Silicon Valley Bank's 2022 annual report called these investor dependent loans: credits whose repayment often relies on the borrower raising more from venture capitalists, or on a sale or IPO. Horizon Technology Finance's annual report defines venture lending by four features, which together make the bet tolerable:

The equity cushion is thick. Runway Growth Finance reports that its portfolio companies had raised about $157 million of equity on average at origination, against an average commitment of about $33 million, so the lender sits senior to roughly five times its own money. Venture lenders therefore spend less time trying to value a company with no profits than judging whether insiders will back another round.

Term Loans, Tranches and Growth Lines

Most venture debt is a growth capital term loan: a committed amount, partly drawn at closing, with the rest available over a draw period of up to two years, sometimes only once the company hits a revenue target or closes its next round. Revolving lines are sized against receivables, billings or contracted revenue, which for a software borrower can mean a line that grows with annual recurring revenue. With no EBITDA to multiply, sizing follows rules of thumb, and SVB's guidance for founders gives three:

Interest-Only Periods, Rates and Fees

A venture loan usually opens with an interest-only period, then amortizes in equal monthly installments to maturity. Rates are mostly floating, a spread over the prime rate or SOFR with a floor; Hercules Capital's 2025 annual report shows almost all of its loans at floating rates.

The ranges come from the 2025 annual reports of Hercules Capital, Trinity Capital, Horizon Technology Finance and TriplePoint Venture Growth; banks usually price below them with tighter terms. The gap between stated rate and all-in yield is the fee stack: a facility fee at closing, prepayment and sometimes success fees, and an end-of-term payment, a fixed percentage of the original principal due in cash at maturity or on any prepayment, which lifts the yield without raising the monthly rate.

How Warrant Coverage Works

Warrant Coverage

Warrant coverage is the size of the equity warrant a lender receives with a venture loan, expressed as a percentage of the loan amount. Ten percent coverage on a $10 million loan lets the lender buy $1 million of shares at a set exercise price, usually the latest round price, so it profits only if the shares rise above that strike.

Hercules Capital's 2025 annual report puts its warrant coverage at roughly 3% to 20% of principal, with the strike generally at the most recent round price; TriplePoint Venture Growth cites 2% to 10%. The share count follows directly:

On a $10 million loan with 10% coverage and a $5 Series A price, that is 200,000 shares, worth only their option value above $5, not $1 million. For the borrower, part of the loan proceeds is allocated to the warrant, creating a debt discount that lifts the effective interest rate above the coupon, and the shares enter the diluted count under the same treasury stock logic used for warrants and convertibles. The payoff is a portfolio game: Silicon Valley Bank held warrants in 3,234 companies at the end of 2022, and 65 positions worth more than $1 million each made up about 52% of the portfolio's fair value.

Light Financial Covenants, Sharp Triggers

A loss-making startup would fail a leverage or interest coverage test on day one, so venture loans skip most ratio covenants. Runway Growth Finance's annual report says venture lenders focus instead on a short list of triggers, a package that would count as covenant-lite by traditional standards:

Material Adverse Change (MAC) Clause

A material adverse change clause lets a lender declare a default, or decline to advance further funds, if the borrower's business, finances or prospects deteriorate significantly. It is deliberately vague, which gives the lender discretion, and in venture debt it usually sits beside an investor abandonment clause.

That discretion has teeth. Hercules Capital's filings say its credit agreements relieve it of undrawn commitments when a borrower suffers materially adverse events, so the second tranche a founder treats as insurance can vanish exactly when it is needed.

Security Over Everything, Including IP

Venture loans are senior secured, usually with a first lien on all assets. Because the valuable asset is often intellectual property, lenders either take a lien on it or a negative pledge barring the company from pledging it to anyone else, approaches Horizon Technology Finance and Hercules Capital both describe. Bank lenders also usually require the company's operating accounts to sit with them.

Banks Versus Specialist Lenders

Venture banks lend most cheaply because the loan is one product in a relationship that includes deposits and payments; Silicon Valley Bank, now a division of First Citizens Bank, and HSBC's innovation banking team are examples. Specialist non-bank lenders charge more. Several are business development companies, the vehicles that also fund much of private credit and direct lending: Hercules Capital, Trinity Capital, Horizon Technology Finance, TriplePoint Venture Growth and Runway Growth Finance all describe lending to venture- or institutionally backed growth companies as their core business, and Hercules says it has committed more than $25 billion to over 700 companies since inception. Private venture debt funds complete the field.

Silicon Valley Bank and the Market Since

The Federal Reserve's review of SVB's failure notes that the bank said it served nearly half of US venture-backed technology and life sciences companies, and that about 94% of its deposits were uninsured. Depositors pulled more than $40 billion on March 9, 2023, and regulators closed the bank the next day. For borrowers, the episode exposed a concentration risk: the loan and the payroll cash often sat in the same bank. First Citizens later bought SVB's loans and deposits from the FDIC, and HSBC, having bought SVB's UK arm for £1, built a US innovation banking team of more than 40 bankers led by former SVB executives.

Lending then fell and rebounded on bigger deals, per the PitchBook-NVCA Venture Monitor, whose figures for past years are revised as late deals are logged:

The first three quarters of 2026 already beat every full year on value, with fewer than half of 2025's deals, as very large AI loans pushed the average up. The Monitor's year-end 2024 edition linked much of that year's jump to AI, with more than 36% of new debt going to AI projects and skewed toward outsized deals such as CoreWeave's $7.5 billion raise, and noted that early-stage lending had not fully recovered. A separate 2025-2026 Venture Debt Review from Runway Growth Capital and PitchBook also called 2025 a record, at $68.8 billion across roughly 1,000 deals, with a median of about $5.5 million. The datasets are cut differently, so the median is the better guide to a typical loan.

The Two Offers

Ferrovane has 18 million fully diluted shares priced at $5 in its Series A, so a flat round values it at $90 million pre-money, and the founders own 40%. Every number is invented and rounded; at about 11% of the valuation, the loan is deliberately on the large side.

The loan costs about $2.5 million of interest plus $500,000 of fees, roughly $3 million in all. Runway is where it disappoints: debt service starts at once, and principal payments of about $417,000 a month begin just as cash runs low, so $10 million of debt buys about six extra months rather than ten.

Where the Answer Flips

If Ferrovane hits its milestone and later sells for $300 million, Option A's new shares collect a tenth, or $30 million, a $20 million gain on the $10 million they put in. Under Option B, each of 18.2 million shares is worth about $16.50, so the warrants are worth roughly $2.3 million after the $5 strike; with $3 million of interest and fees, the loan cost a little over $5 million, roughly a quarter of the equity's cost in the good outcome.

If the Series B never comes and Ferrovane sells for $15 million sixteen months in, about $8.3 million of principal plus the $400,000 end-of-term payment and any prepayment fee come off the top: the lender takes close to $9 million before any shareholder is paid. Under Option A, all $15 million would have reached shareholders. Both cases ignore the liquidation preference stack, which sits behind the lender and ahead of common and makes the bad outcome worse still for founders.

Venture debt versus equity is a favorite test of financing judgment: Work through dilution, warrant and capital structure questions with worked answers, start practicing interview questions for free and see which trade-offs you can explain under pressure.

Where It Works

The NBER study found that startups using venture debt were less likely to close and somewhat more likely to reach an IPO or acquisition; the authors attribute most of the survival difference to selection, since lenders pick stronger companies, but still find a positive effect after controlling for it. In practice the loan pays for itself when a specific milestone, worth more to the next round's price than the loan costs, sits comfortably inside the extra months, and when missing it would not turn the debt into a crisis.

Where It Backfires

The classic mistake is borrowing to delay an inevitable down round. If the milestone is not really within reach, the loan only postpones the repricing, and the company meets the lower valuation later with less cash, amortization draining several hundred thousand dollars a month, and a creditor ranking ahead of new investors who know part of their money will repay the lender.

Two traps follow. In a covenant spiral, a minimum liquidity or performance-to-plan test trips just as the company is raising, and a default notice is the last thing a lead investor wants to see. Then comes lender control: once in default, a senior secured lender can refuse draws, demand repayment and push for a quick sale that clears the loan, even if a longer process would have paid shareholders more.

Financing trade-offs like venture debt versus equity run through technical interviews: The 160-page PDF covers valuation, capital structure and dilution alongside the core technical frameworks, access the IB Interview Guide and see where venture lending fits.

On Live Deals

On a sell-side process for a venture-backed company, the loan is usually repaid at closing with any prepayment fee and the end-of-term payment, the warrants become a line in the payout waterfall, and the lender's payoff letter is a closing deliverable; in an IPO, outstanding loans can be repaid from the proceeds. A growth equity fund leading a late-stage round has to price around any loan ranking ahead of its new preferred, one reason the differences between growth equity, buyout and venture capital matter in practice. Venture lending teams at banks and BDCs are also a niche private credit career path.

In Interviews

Venture debt questions test seniority and dilution together. Two come up most:

Venture debt stops looking like a contradiction once the underwriting is clear. The lender is betting that equity investors will keep funding a company they have just backed, and every term protects that bet: seniority, fast amortization, liquidity and investor abandonment triggers, and warrants that share a little in the winners.

In PitchBook's data the label stretches from a loan of a few million dollars to a Series A software company all the way to multi-billion-dollar AI infrastructure financings, so the first question in any conversation about the market is which kind of deal is meant.

For candidates, the topic is a compact test of capital structure judgment. Explain who gets paid first, what a warrant is really worth, and whether the extra months will change the next round's price, and you have shown an interviewer that you can read both sides of a startup balance sheet.